MathGroup Archive 1995

[Date Index] [Thread Index] [Author Index]

Search the Archive

[Q] Need Package for ANOVA

  • To: mathgroup at
  • Subject: [mg1699] [Q] Need Package for ANOVA
  • From: dave at (David Johnson)
  • Date: Mon, 17 Jul 1995 03:24:28 -0400
  • Organization: Northwestern Unviversity


Sorry if this is an iteration of an earlier post.

I have a sets of data that are roughly 8x4000 in matrix form.  
In other words we have many sets of data that each contain 
8 populations of 4000 samples each.

Our goal is to first prove that each experiment of 8 populations are similar.
Following that we want to extend this idea and show that the experiments are

I know this has been done and is availiable in packages such as
StatView but they have the unfortunate disadvantage of requiring the user
to manually input (cut/copy/open) each datafile.  This is not practicle for
us because we have hundreds of these to do.  

I want to run an ANOVA on these and I'd like to find a pre-written
package that has an ANOVA function that can deal with a large matrix
and then output a report similar to Regress[].  Ideally the package might 
have options for post hoc tests (ie Fisher, Dunn), although this is 
not absolutely necessary for our current data sets.  Can anyone give me a 
little direction where I can find such a package?
We'd be willing to purchase it if it is commercially availiable.

I've looked through MathSource, the Mathematical Journal and whatever
books on Mathematica I've been able to easily obtain and I haven't seen
something like this. 

..Thank's in advance, 


David A. Johnson (office)= (312) 503-0408     (pager) = (312) 820-3469
Research Engineer
Feinberg Cardiovascular Research Institute
Northwestern University
Tarry 12-729
310 E. Superior St.
Chicago, Il.  60611   

  • Prev by Date: Looking to buy Mathmatica for Windows?
  • Next by Date: Re: Problems with NIntegrate
  • Previous by thread: Looking to buy Mathmatica for Windows?
  • Next by thread: FindRoot: Slight variation of usage